It has been observed that reduced metabolic rates and high 24-hour respiratory quotients are predictors of subsequent weight gain in Pima Indians. It is unknown whether of not this is true for other populations and whether the fasting RQ is a predictor of weight gain as well. Seven- hundred and seventy-five men (aged 18-98 years) participating in the Baltimore Longitudinal Study of Aging were followed for an average of 10 years. Basal metabolic rate and RQ were measured on their first visit and related to subsequent weight change. Deviations from the predicted value of basal metabolic rates (predicted from their estimated fat free mass) were calculated. Average weight change was 0.1 kg (s.d.6.4 kg); 122 (15%) gained more than 5 kg and 40 (5%) more than 10 kg during the follow-up. After adjustment for initial age, body mass index, fat free mass, and duration of follow-up, RQ, but not BMR or deviations from predicted BMR, was positively related to weight change (p less than 0.001). Major weight gain (5-15 kg) was related to initial RQ in non-obese men only (body mass index less than 25 kg/m2). From Cox proportional hazard regression analyses the adjusted relative risk of gaining 5 kg or more in initially non-obese men with a fasting RQ of 0.85 or more was calculated to be 2.4 (95% confidence interval: 1.1-5.3) compared to men with a fasting RQ less than 0.76. Conclusion: A relatively high fasting RQ is a weak but significant predictor of substantial weight gain in non-obese white men.